How can Conversational Agents be leveraged to effectively engage with patients?

Video Abstract

ETH PhD Candidates Theresa Schachner and Christoph Seraphin Gross, together with Andrea Hasl (International Max Planck Research School on the Life Course), Prof. Dr. Tobias Kowatsch (ETH Zurich & University of St.Gallen), and Prof. Dr. Florian von Wangenheim (ETH Zurich) have published a paper on interaction styles in conversational agents used in medical settings.

In the paper published in JMIR, the authors explore how interaction styles adopted by medical professionals can be implemented in conversational agents. The authors pose the question of how different interaction styles can be identified and used in practice, focusing on differences between paternalistic and deliberate interaction styles. The authors aimed to create a systematic approach to developing and implementing specific interaction styles in a conversational agent, as well as to find out if patients were able to identify the conversational style used in their care.

To explore this, the authors developed a comprehensive mechanism to induce either a deliberative or a paternalistic interaction style into an interaction between a patient with Chronic Obstructive Pulmonary Disease (COPD) and a conversational agent. A between-subjects web-based experiment was used in which participants were randomly assigned to engage with a conversational agent with one of the two distinct styles. The study’s results demonstrated that in more than 80% of cases, patients were able to identify whether their conversational agent was using a paternalistic or a deliberate interaction style.

The authors hope that their research will help advance the use of personalised conversational agents and encourage research into matching specific interaction styles to different conditions, including care settings and patient groups.

For further details, please refer to the published article.

Share this post